Skip to main content
Version: 2.0

Corrects hallucinations in generated text based on source documents

POST 
/v2/hallucination_correctors/correct_hallucinations

This endpoint identifies information in generated text that is not supported by the provided source documents and offers corrections with minimal changes. This can be used standalone or as part of a RAG workflow where the HHEM score indicates potential hallucinations.

Request​

Responses​

Successfully analyzed the text for hallucinations